基于等价关系的关联规则的挖掘

由于在实际的数据挖掘过程中容易出现无用的频集和冗余的规则，所以降低频集和规则的冗余度可大大提高挖掘的质量，这也是数据挖掘中一直被关注的问题，提出了一个用等价类生成关联规则的方法，算法主要在频集的基础上建立项集的等价关系，进而对项集划分等价类，同时将得到的关联规则划分为精确关联规则和近似关联规则两个集合，通过等价类，不但可以很容易地生成所需要的关系规则，同时可以方便地判断数据之间依赖关系的强弱，同时，项集的等价关系在实际应用中也很有利用价值，算法最后的规则结果集剔除了由来自同一等价类中的面集的重复出现构成的冗余规则，从而得到了较小的关联规则集合。

Equivalence relation based association rules mining

The usefulness set for discovered association rules is important in data mining.To reduce the redundancy of frequent itemsets and association rules could improve the quality of mining.In this paper,we propose a new algorithm based on equivalence relation to reduce the frequent itemsets and association rules and depart the rules to the set of exact rules and the set of approximate rules.The items are classified into equivalence classes.After the classification,the discovering process for rules becomes easier.The equivalence class is valuable in practice.Our approach provides a small set of rules without redundancy of itemsets in the same equivalence class.